HOW TO DESIGN USABLE
SYSTEMS
by John D. Gould
presented by
Andrew Trieu
Subha Narasimhan
ICS 205 - Human Computer Interaction
Department of ICS at UCI
May 15, 2002
Introduction
The focus of this paper is to describe a process of
system design that will help you to design good
computer systems for people.
A good computer system is a system that is:
• easy to learn,
• easy to use,
• contain the right functions, and
• is well liked.
Overview
To design good systems, the author suggests the
following four principles:
Early and continual focus on users - make direct
contact with users and what they do,
Early and continual user testing - users work with
simulations and prototypes early in the project and
measure performances/reactions qualitatively and
quantitatively,
The Four Principles (Con’t)
Iterative design - modify the system according to
results from user test, and the testing cycle is
repeated,
Integrated design - all aspects of usability evolve in
parallel and under one focus.
Overview (Con’t)
The author suggested 20-30 informal methods to carry
out the four principles in this usability design
process.
As a means of showing where these four principles fit
into the system design, the author explains how to
carry out the steps (i.e.the four principles) in four
phases.
Usability Design Phases
The author explains how to carry out the steps (i.e. Four
Principles) in four phases.
 Gearing-Up Phase - information gathering and
conceptualization phase.
 Initial Design Phase - early focus on users.
Specification is based on existing system, etc. and
information about workers and their environment is
collected.
 Iterative Development Phase - modification and
evaluation is ongoing based on user feedback.
 System Installation Phase - installing & using the
system by the client.
Usability Has Many Aspects
There are many aspects to usability which must be
taken into account if a system is considered to be
good.
The components of usability are:
 System Performance
 System Functions
 User Interfaces
 Reading Materials
Components of Usability (Con’t)







Language Translation
Outreach Program
Ability for customers to modify and extend
Installation
Field Maintenance and Serviceability
Advertising
Support-group users
Principle 1. Early And Continual
Focus On Users
A first step in designing a system is to decide:
 who the users will be and
 what they will be doing with the system.
This is done either before starting to design the system,
or at an early stage after obtaining some general ideas
about it.
Designers Often Avoid This: the authors observed two
serious types of reluctance:
 one a reluctance to define the users, and
 the other a reluctance to take the definition seriously.
Methods To Carry Out
Principle 1
Talk with Users. Talk with the intended users. Ask
them about their problems, difficulties, wishes, and
what works well now. After analyzing their
response, tell them what you have in mind and get
their reactions.
Visit Customer Locations. Visit potential locations for
your system, particularly if these environments are
new to you.
Methods To Carry Out
Principle 1 (Con’t)
Observe Users Working. Visit the workplaces of users.
Observe potential users doing their jobs. You may
learn what users do when they don’t know something
or get in trouble. This should help you in designing
your outreach program. See what they have
difficulty with and what they dislike. The
implications of these findings must then be
incorporated into the design of your system.
Methods To Carry Out
Principle 1 (Con’t)
Videotape Users Working. Make a videotape of a few
users working and show it to other members of the
design team. Brief videotapes of users having
difficulty carry more punch than do tables with lots
of numbers.
Learn about the Work Organization. Here the emphasis
is upon understanding the organization of the work
that your system is intended to help. This is
particularly important where many different user
groups are involved.
Methods To Carry Out
Principle 1 (Con’t)
Thinking Aloud. Having potential users think aloud as
they are doing their actual work. Doing this while
actually performing their job may yield different
insights than having them reflect on their work later.
Try It Yourself. Sometimes it can be rewarding to try a
worker’s job yourself.
Participative Design. Make intended users part of the
design team. Experienced workers know a lot more
about what they do than do drop-in designers. Also,
they want to be actively involved in their work, and
in the decisions that affect their work lives.
Methods To Carry Out
Principle 1 (Con’t)
Expert on Design Team. This expert is a consultant and
not a full-fledged member of the design team.
Putting experts on the design team is simple another
way to phrase “participative design”.
Task Analysis. Task analysis is an analytical process
used to determine the specific behaviors required of
people in a man-machine system. Task analysis is a
comparison between the demands the task places on
the human operator and the capabilities of the
operator to deal with them. It is usually carried out
through observation, interviews, or questionnaires.
Methods To Carry Out
Principle 1 (Con’t)
Surveys and Questionnaires. The data obtained from
surveys and questionnaires can be useful.
Testable Behavioral Target Goals. Most new systems
specify in advance physical performance and
capacity targets. Measurable behavioral targets, and
where your developing system stands with respect to
them, give management a metric to understand what
progress has been made, and what is still required.
Principle 2. Early And Continual
Users Testing
Your job is to design a system that works and has the
right functions so that users can do the right things.
You won’t know whether it is working right until you
start testing it.
From the very beginning of the development process,
and throughout it, intended users should carry out
real work using early versions of training materials
and manuals, simulations and prototypes of user
interfaces, help systems, and so forth. The emphasis
here is upon measurement, informal and formal. If
you measure and then make appropriate changes, you
can hill-climb toward an increasingly better system.
Methods To Carry Out
Principle 2
Printed or Video Scenarios. As a starting point, sketch
out a few user scenarios on paper and show them to
members of the design team. Provide exact details
exactly what keys users must press and the response
of the system. Then carry the process one step
further by typing up these scenarios to be shown to
the prospective users for their reaction. What you
have done with this procedure is to identify and
organize functions in a way that intended users can
understand and react to. You are designing the
system from the user’s point of view.
Methods To Carry Out
Principle 2 (Con’t)
Early User Manuals. Begin writing the user manual
before any code is written. Intended users can react
to this helpfully, since the system is being described
in the appropriate fashion.
Mock-ups. By having mock-ups, we might identify
situations or problems that we might not have been
otherwise as easily envisioned.
Methods To Carry Out
Principle 2 (Con’t)
Simulations. Much informal and formal
experimentation can be carried out by simulating
important parts of the system. These simulations can
help identified both how people used the systems and
how they felt about them.
Methods To Carry Out
Principle 2 (Con’t)
Early Prototyping. Early prototyping can be made
possible through the use of designer toolkits or user
interface management systems. Try to develop
pieces of your system to the point where potential
users can carry out pre-defined problems. You will
learn things you have possibly missed through
simulations such as effects of multiple simultaneous
users. You can measure people’s performance and
feelings.
Note: prototyping is expensive, but necessary.
Methods To Carry Out
Principle 2 (Con’t)
Early Demonstrations. Demonstrate working pieces of
your system to anyone who will take the time to
watch. It is even better to let users try it themselves
on a brief task. Successful demonstrations of pieces
of your system and manuals give management and
customers confidence that you are making progress.
Thinking Aloud. Performance measures, such as time
and errors, do not give a clear indication of what is
bothering users or what may be the source of an user
error.
Methods To Carry Out
Principle 2 (Con’t)
Make Videotapes. Besides being useful for measuring
time, errors, and user attitudes, brief videotapes of
users attempting to use a new system have
tremendous impact upon management, especially
when the users are having problems.
Methods To Carry Out
Principle 2 (Con’t)
Hallway and Storefront Methodology. By putting a
mock-up, simulation, or an early prototype in an
obvious public place, passers-by just naturally are
attracted to use it. This provides a source of
invaluable comments and surprises. What gets
learned in storefront and hallway methodology is
valuable for user guides, user interfaces, display sizes
and colors, identification of required functions, help
systems, and the design and looks of the workstation.
Methods To Carry Out
Principle 2 (Con’t)
Computer Bulletin Boards, Forums, Networks, and
Conferencing. Existing, extensive computer
networks allow designers to send out a partial or an
entire new user interface and obtain feedback from
users all over the world -- most of whom would
otherwise be unknown to the designer. These users
can provide immediate feedback to the designer on
their preferences, needs, and suggestions for change.
However, one factor that motivated users to provide
feedback is the speed with which we respond to
them.
Methods To Carry Out
Principle 2 (Con’t)
Formal Prototype Test. Much of the emphasis so far has
been upon informal experimental results. However,
the author encourages formal experimentation where
possible. If skilled human factors people are available
to develop the experimental and statistical designs,
then an even more accurate and valuable assessment
can be made. Furthermore, as the project nears its end,
it is better if you can have an outside group do the final
evaluation.
Methods To Carry Out
Principle 2 (Con’t)
Try-to-Destroy-It contests. When the system is near its
end, you might want to turned the system over to a
group of college students and let them try to find
bugs, crash it, break into it, etc. This will give you
information that you might not get from all the
previous methods.
Field Studies. Laboratory and hallway studies go only
so far. Putting your system into the field for a test
can remind you of problems that you have put out of
mind or identifies problems that other methodologies
do not get at.
Methods To Carry Out
Principle 2 (Con’t)
Follow-Up Studies. Once a system has been released,
studying how actual customers use it has value for
subsequent releases and related new products. This
work serves as a validation of the earlier prototyping
and iterative design efforts, and it is particularly
important in assessing the usefulness of various
functions and what new functions are required.
Principle 3. Iterative Design
Methods
System Development Work Organization
Identification of required changes
Willingness to make the changes
Caution: Should be aware of cyclic phases.
Methods [2]
 Software Tools - Toolkits like MAC, UIMS
Provide ability to make changes
Separate User Interfaces from Functionality
Can be a basis of prototyping
Bring user-interface to the control of non-programming
specialists
Reduce cost by making programmers more productive
Enforce ‘Best Practices’ inherently
Facilitate User interface and cross system consistency
Integrated Design
IBM’s Olympic Message System, Field test e.g. from
Boies et al. The word ‘Olympics’ had to be changed to
‘Olympic’.
All User interfaces, help messages required the change.
Messages were in 12 languages. All the speakers had to be
recalled.
User guides required the change and they were in 12
languages too, some with unprintable chars.
Lettering on the signs had to be changed.
Was handled easily because all usability aspects were the
responsibility of one person and all the messages were stored
in one file.
Principle 4. Integrated
Design
All aspects of Usability evolve in parallel.
Projects are managed to only a few goals – low cost,
processing speed, compatibility, schedule,
reliability..Usability should be included as one of these
goals.
Development groups are fraction - Difficult to coordinate and achieve integrated design.
Methods for Integrative Design
Consider all aspects of usability in design.
Requirements for both functionality and
interfaces should be as complete as possible
before design.
All usability aspects should be under one focus
and one group should be allocated the sufficient
resources at the very beginning to drive
usability.
Methods [2]
Assure adequate System reliability, robustness
and responsiveness.
User manuals should be continuously updated.
Outreach program – help support, training,
hotlines, video
Installation, Customization and field
maintenance.
Where should a designer start?
Starting Points
Define the system: The most important starting point is
to define what the system will be. For example, who
will use it, what should it do, and why the users and/or
organization will benefit from it.
Follow-on system: Most computer systems are not new.
They always are new releases, or extensions of already
existing in-house applications.
New influential systems: New technologies are often
the driving force to create new systems.
Starting Points [2]
User circumstances: Some times a good starting point
is to build on existing user knowledge, skills, and
resources.
Journals, proceedings, demonstrations: Paging through
journals and proceedings can be a source of starting
ideas.
Other designers and consultants: Other designers and
consultants usually are benefit for you.
Summary
Usability is combination of many factors, each of
which is often developed independently.
User Interfaces are becoming a large part of the code
and existing guidelines are not enough.
Usability should be considered from the very beginning
and throughout the development phase and beyond too.
Usability should be measured and iterated
continuously.
Conclusion [1]
The paper described:
 The components of usability.
 Four principles to guide usability design process.
Early continual focus on users; make direct contact with
users and what they do.
Early and continual user testing ; users work with
simulations and prototypes early in the project and measure
performances/reactions qualitatively and quantitatively.
Iterative design; modify the system according to results
from user test.
Integrated design; all aspects of usability evolve in parallel
and under one focus.
Conclusion [2]
 Methods for achieving the above four principles.
 Four Phases of design process to fit the principles in :
Gearing up Phase - Information gathering and
conceptualization phase
Initial Design Phase - Early focus on users. Specs are based
on existing systems, etc. and info about workers and their
environments is collected.
Iterative Development Phase - Modification and
evaluation is ongoing based on users feedback
System Installation Phase- Installing and using the system
by the client.
 Starting points for a designer.